The goal of this study is to identify molecular markers (gene expression levels and germline polymorphisms) associated with clinical outcome and toxicity among Dukes' stage B2, B3 or C rectal cancer patients treated with fluorouracil (5-FU) based chemotherapy and radiation. Failure of adjuvant therapy represents a major problem in rectal cancer, and acute gastrointestinal and hematologic toxic effects are considerable and can be life-threatening. Establishing associations between molecular markers of drug resistance, treatment response, and clinical toxicity, may ultimately result in more successful and less toxic chemotherapeutic regimens for rectal cancer patients. As part of initial studies, we have shown that intratumoral gene expression (e.g., dipyrimidine dehydrogenase) and germline polymorphisms (e.g., thymidylate synthase, TS) can predict response and overall survival in patients with metastatic colon cancer treated with 5-FU. We propose to focus on markers in several key pathways, including 5-FU and folate metabolism, DNA repair, and angiogenesis. Among 788 patients with rectal cancer registered to a collaborative Phase III study of rectal cancer (INT-0144, SWOG 9304) we will determine 1) whether intratumoral gene expression levels (RNA) of enzymes involved in folate and 5-FU metabolism, DNA repair, and growth factors, are associated with clinical outcome; and 2) whether germline polymorphisms (DNA) of genes encoding proteins involved in 5-FU and folate metabolism, detoxification and DNA repair, or angiogenesis, are associated with toxicity and clinical outcome. The study uses a large clinical trial patient population with excellent assessment of outcomes and toxicities, and utilizes existing specimens for a cost-effective approach. Results from this study may aid in the development of future treatment strategies for patients with rectal cancer.